Protestant Epistemology as Juridical Contestation: How Evangelical Institution-Building Contests Colonial Legal Frameworks

Here's an updated README description that incorporates both analyses: Analyzing Religious and Demographic Change in Latin America with Google Colab This repository contains Python scripts for analyzing interconnected trends in Catholic affiliation decline and household size changes across Latin America, with particular focus on Puerto Rico. The analyses utilize descriptive statistics, growth rate analysis, K-means clustering, and regression modeling to explore institutional factors driving religious and demographic transformation during the colonial period (1910-2014). Repository Contents 1. Catholic Affiliation Analysis (Protestant.ipynb) Analyzes Catholic decline across 19 Latin American countries (1910-2014), demonstrating that U.S. colonial influence is the strongest predictor of religious transformation. Key Findings: Countries with high U.S. institutional presence experienced Catholic decline averaging 37.0 percentage points (1970-2014) vs. 18.2 points in low-influence countries Statistical significance: p < 0.001, R² = 0.559, Cohen's d = -2.21 Puerto Rico declined from 87% Catholic (1970) to 56% (2014)—a 31-point drop coinciding with Protestant missionary expansion, Operation Bootstrap industrialization, and eugenic sterilization campaigns 2. Household Size Analysis (Family.ipynb) Examines Puerto Rican household size trends using IPUMS census microdata (1910-2000), revealing demographic collapse paralleling religious transformation. Key Findings: Household size declined 33% over 90 years (4.89 → 3.26 persons) Post-1970 acceleration: decline rate 2.4x faster than pre-1970 period Temporal correlation with Catholic decline and sterilization campaign (1930s-1970s) Evidence supporting "coordinated institutional assault" thesis Data Sources Catholic Affiliation Data: Catholic_Affiliation_in_Latin_America.xlsx: Country-level religious affiliation (1910, 1950, 1970, 2014) Source: Pew Research Center, national censuses, historical church records Household Size Data: Puerto_Rican_family_size_census_data.xlsx: Average household size from Puerto Rico census samples Source: IPUMS USA (Integrated Public Use Microdata Series), University of Minnesota Census years: 1910, 1920, 1930, 1970, 1980, 1990, 2000 Sample sizes: 6,649 to 55,037 households per year Running the Analysis in Google Colab Setup Instructions Open the Notebooks in Google Colab: Click on Protestant.ipynb (Catholic analysis) or Family.ipynb (household size analysis) Click "Open in Colab" or upload manually to Google Colab Upload the Data Files: Download the corresponding Excel file from this repository: Catholic_Affiliation_in_Latin_America.xlsx for Protestant.ipynb Puerto_Rican_family_size_census_data.xlsx for Family.ipynb In Google Colab, use the file explorer (left sidebar) to upload the Excel file to /content/ Modify the Filepath: Locate the code block defining the filepath: if __name__ == "__main__": filepath = '/content/Catholic Affiliation in Latin America.xlsx' # Update this df = main(filepath) Verify the filepath matches your uploaded file location (check the Files panel) Run the Notebook: Click "Connect" to start a runtime Select "Runtime" → "Run all" to execute all code blocks sequentially Monitor output as analyses run and visualizations generate Inspect Results: Protestant.ipynb outputs: Descriptive statistics for Catholic affiliation by year Period-specific decline rates K-means clustering analysis (Severe/Moderate/Stable decline groups) Regression results showing U.S. influence effects 4-panel visualization: catholic_decline_analysis.png Family.ipynb outputs: Puerto Rico household size trends (1910-2000) Period-by-period decline analysis Pre-1970 vs. post-1970 acceleration comparison 2-panel visualization: puerto_rico_household_size_analysis.png Download Visualizations: Click the Files icon (left sidebar) Locate output PNG files Right-click → "Download" Methodology: IPUMS Household Size Data Extraction Puerto Rico household size data were extracted from the Integrated Public Use Microdata Series (IPUMS USA), maintained by the University of Minnesota Population Center. We selected census samples spanning 1910-2000, including decennial censuses (1910, 1920, 1930, 1970, 1980, 1990, 2000). Variables extracted included YEAR (census year), STATEFIP (state FIPS code, filtered to 72 for Puerto Rico), NUMPREC (number of persons in household), HHWT (household weight for population estimates), and SERIAL (household identifier). To calculate average household size by year, we collapsed person-level data to household level by selecting unique YEAR-SERIAL combinations, then computed weighted averages using the formula: Average Household Size = Σ(NUMPREC × HHWT) / ΣHHWT for each census year. Data cleaning excluded records with missing critical values and limited household sizes to 1-20 persons to eliminate data errors. The final analysis covers seven census years with sample sizes ranging from 6,649 households (1970) to 55,037 households (1990), providing statistically robust estimates across 90 years of Puerto Rican demographic history. Integration: The "Coordinated Institutional Assault" Thesis These analyses provide complementary evidence for understanding U.S. colonial governance in Puerto Rico as operating through coordinated intervention across multiple domains: Religious transformation: Protestant institutional expansion displacing Catholic dominance (1899-2014) Demographic disruption: Eugenic sterilization campaign affecting 1/3 of women of childbearing age (1930s-1970s) Family structure collapse: Household size declining 33%, accelerating post-1970 Economic transformation: Operation Bootstrap industrialization and agricultural displacement The temporal overlap and institutional coordination among these initiatives suggest systematic colonial biopolitical governance rather than isolated sectoral policies. Technical Requirements Python Libraries: pandas (data manipulation) numpy (numerical computing) matplotlib, seaborn (visualization) scipy (statistical analysis) scikit-learn (clustering, regression) statsmodels (econometric modeling) openpyxl (Excel file handling) All libraries are pre-installed in Google Colab—no local installation required. Citation Data Sources: Pew Research Center (2014). Religion in Latin America: Widespread Change in a Historically Catholic Region. Steven Ruggles, Sarah Flood, et al. IPUMS USA: Version 15.0 [dataset]. Minneapolis, MN: IPUMS, 2025. https://doi.org/10.18128/D010.V15.0 For Academic Use: If using these analyses in research, please cite the underlying data sources and note that analytical code was developed with assistance from Claude (Anthropic) for data processing and statistical computation. Support For questions about: IPUMS data access: https://usa.ipums.org/ Pew Research methodology: https://www.pewresearch.org/ Google Colab usage: https://colab.research.google.com/ Python libraries: Refer to pandas, scikit-learn, statsmodels, matplotlib, seaborn documentation This toolkit enables econometric and demographic analysis without advanced coding skills or local software installation. The integrated analyses reveal how colonial governance operated through coordinated religious, economic, and biopolitical mechanisms to transform Puerto Rican society.# 使用Google Colab分析拉丁美洲的宗教与人口结构变迁 本仓库包含用于分析拉丁美洲地区天主教信徒占比下降与家庭规模变化这两大关联趋势的Python脚本，研究重点聚焦波多黎各（Puerto Rico）。本次分析采用描述性统计、增长率分析、K-means聚类（K-means clustering）以及回归建模等方法，探究1910年至2014年殖民时期推动宗教与人口结构转型的制度性因素。 ## 仓库内容 1. 天主教信徒占比分析（Protestant.ipynb） 分析1910年至2014年间19个拉丁美洲国家的天主教信徒占比下降情况，证明美国的制度性影响力是宗教转型的最强预测因子。 ### 关键研究发现： - 美国制度性影响力较高的国家，其天主教信徒占比在1970年至2014年间平均下降37.0个百分点，而低影响力国家仅下降18.2个百分点 - 统计学显著性：p < 0.001，决定系数R² = 0.559，科恩d值（Cohen's d）= -2.21 - 波多黎各的天主教信徒占比从1970年的87%下降至2014年的56%，降幅达31个百分点，这一变化与新教传教活动扩张、"bootstrap工业化计划"（Operation Bootstrap）以及优生绝育运动同期发生 2. 家庭规模分析（Family.ipynb） 使用IPUMS美国整合公共使用微数据系列（IPUMS USA）的普查微观数据（1910年至2000年）分析波多黎各的家庭规模变化趋势，揭示出与宗教转型同步的人口结构崩塌。 ### 关键研究发现： - 90年间家庭规模下降33%，从每户4.89人降至3.26人 - 1970年后下降速度加快：降幅是1970年前的2.4倍 - 与天主教信徒占比下降及绝育运动（1930年代至1970年代）存在时间相关性 - 为"协同制度性冲击"论点提供了实证支撑 ## 数据来源 ### 天主教信徒占比数据： - Catholic_Affiliation_in_Latin_America.xlsx：分国家的宗教信徒占比数据（1910、1950、1970、2014年） 数据来源：皮尤研究中心（Pew Research Center）、全国普查数据及历史教会档案 ### 家庭规模数据： - Puerto_Rican_family_size_census_data.xlsx：波多黎各普查样本的平均家庭规模数据 数据来源：明尼苏达大学整合公共使用微数据系列（IPUMS USA） 普查年份：1910、1920、1930、1970、1980、1990、2000年 样本量：每年6649户至55037户不等 ## 在Google Colab中运行分析 ### 设置步骤： 1. 在Google Colab中打开笔记本： - 点击Protestant.ipynb（天主教分析）或Family.ipynb（家庭规模分析） - 点击"在Colab中打开"，或手动上传至Google Colab 2. 上传数据文件： - 从本仓库下载对应Excel文件： - 针对Protestant.ipynb，下载Catholic_Affiliation_in_Latin_America.xlsx - 针对Family.ipynb，下载Puerto_Rican_family_size_census_data.xlsx - 在Google Colab中，使用左侧边栏的文件浏览器将Excel文件上传至`/content/`目录 3. 修改文件路径： - 找到定义文件路径的代码块： python if __name__ == "__main__": filepath = '/content/Catholic Affiliation in Latin America.xlsx' # Update this df = main(filepath) - 确认文件路径与你上传的文件位置一致（可查看文件面板确认） 4. 运行笔记本： - 点击"连接"以启动运行时环境 - 选择"运行时"→"全部运行"以按顺序执行所有代码块 - 监控输出以查看分析运行结果与可视化图表生成情况 5. 查看结果： - Protestant.ipynb的输出包括： - 各年份天主教信徒占比的描述性统计 - 分阶段下降速率 - K-means聚类分析（严重/中等/稳定下降组别） - 反映美国影响力效应的回归结果 - 四面板可视化图表：`catholic_decline_analysis.png` - Family.ipynb的输出包括： - 1910年至2000年波多黎各家庭规模变化趋势 - 分阶段下降分析 - 1970年前与1970年后的下降速率对比 - 双面板可视化图表：`puerto_rico_household_size_analysis.png` 6. 下载可视化图表： - 点击左侧边栏的文件图标 - 找到输出的PNG文件 - 右键点击→"下载" ## 方法论：IPUMS家庭规模数据提取 波多黎各的家庭规模数据提取自明尼苏达大学人口中心维护的整合公共使用微数据系列（IPUMS USA）。我们选取了1910年至2000年的普查样本，包括十年一次的普查数据（1910、1920、1930、1970、1980、1990、2000年）。提取的变量包括：YEAR（普查年份）、STATEFIP（州联邦信息处理标准代码，筛选为波多黎各的72号代码）、NUMPREC（家庭人数）、HHWT（用于人口估计的家庭权重）以及SERIAL（家庭标识符）。 为计算各年份的平均家庭规模，我们通过选取唯一的YEAR-SERIAL组合将个人层面的数据汇总至家庭层面，随后使用以下公式计算加权平均值：平均家庭规模 = Σ(NUMPREC × HHWT) / ΣHHWT，其中求和范围为各普查年份的所有家庭。数据清洗环节剔除了存在关键缺失值的记录，并将家庭人数限定在1至20人之间以排除数据错误。最终分析覆盖7个普查年份，样本量从6649户（1970年）至55037户（1990年）不等，为波多黎各90年的人口历史提供了统计上稳健的估计结果。 ## 整合分析："协同制度性冲击"论点 本研究的两项分析为理解美国对波多黎各的殖民治理提供了互补证据，证明其通过多领域协同干预实施治理： - 宗教转型：新教制度扩张取代天主教的主导地位（1899年至2014年） - 人口结构破坏：优生绝育运动影响了1/3的育龄女性（1930年代至1970年代） - 家庭结构崩塌：家庭规模下降33%，1970年后下降速度加快 - 经济转型：bootstrap工业化计划（Operation Bootstrap）工业化进程与农业人口流失 这些举措之间存在时间重叠与制度协同，表明其背后是系统性的殖民生命政治治理，而非孤立的部门政策。 ## 技术要求 Python库： - pandas（数据处理） - numpy（数值计算） - matplotlib、seaborn（可视化） - scipy（统计分析） - scikit-learn（聚类、回归） - statsmodels（计量经济学建模） - openpyxl（Excel文件处理） 所有库均已预装在Google Colab中，无需本地安装。 ## 引用说明 ### 数据来源： - 皮尤研究中心（2014）. 《拉丁美洲的宗教：历史天主教地区的广泛变迁》 - Steven Ruggles、Sarah Flood等. IPUMS USA：版本15.0 [数据集]. 明尼阿波利斯，明尼苏达州：IPUMS，2025. https://doi.org/10.18128/D010.V15.0 ### 学术引用： 若在研究中使用本分析，请引用上述基础数据来源，并注明分析代码由Anthropic的Claude协助开发以完成数据处理与统计计算。 ## 支持与帮助 若有相关疑问： - IPUMS数据访问：https://usa.ipums.org/ - 皮尤研究中心方法论：https://www.pewresearch.org/ - Google Colab使用方法：https://colab.research.google.com/ - Python库参考：请查阅pandas、scikit-learn、statsmodels、matplotlib、seaborn的官方文档 本工具包无需高级编程技能或本地软件安装即可开展计量经济学与人口结构分析。整合后的分析揭示了殖民治理如何通过协同的宗教、经济与生命政治机制，重塑波多黎各的社会结构。